CVE-2026-111392 — SGLang through 0.5.21 KV Cache Poisoning via mm_hashes in /generate
SGLang through 0.5.21 trusts caller-supplied mm_hashes values as the multimodal prefix-cache key, allowing remote attackers to poison cached KV entries across requests. Attackers can send a different image under a known or predicted hash via /generate to disclose another request's image content or corrupt victims' responses.
Who exploits it and how urgent it is
A 0–100 score combining confirmed exploitation, ransomware and APT use, public exploits, EPSS and severity.
How the score was calculated
- Technical severity (medium)+3
- CISA SSVC: exploitation "poc"+5
The score orders the patch queue; it does not replace your asset context (internet exposure, criticality, compensating controls).
Triage
Is it exploited, how likely is exploitation, what does it touch, and how severe do the sources call it.
CISA SSVC assessment CISA
Three decision points CISA publishes for the CVEs it assesses (SSVC 2.0.3). A stakeholder decision, not a severity score.
Remediation
Fixes and mitigations documented by the sources — linked to the original advisory.
No fix documented by the sources we poll. That does not mean none exists: check the vendor advisory.
First 24 hours
Ordered from the record’s own fields — exposure first, because you cannot patch what you have not found. A rule-generated starting point.
- Identify exposed assets running the affected vendor/product/version combinations (Sgl-project SGLang).
- The attack is network-based and needs no privileges: restrict exposure (ACLs, VPN, WAF) until patched.
- No fix documented in the sources: reduce exposure and monitor the vendor advisory.
- Search available logs for exploit probes matching the vulnerability class (integrity), errors and authentication anomalies.
Affected scope
Vendor, product and versions as the sources word them.
| Vendor | Product | Versions | Status | Source |
|---|---|---|---|---|
| Sgl-project | SGLang | ≥ 0.5.13 and ≤ 0.5.21 | Vulnerable | CVE Program (cve.org) e CISA ADP |
Attack characteristics
The CVSS vector, decoded. It describes the attack, not your exposure to it.
Every base score collected
Sources score independently and may disagree; each row says who scored it and under which version.
| Score | Version | Severity | Expl. | Impact | Scored by |
|---|---|---|---|---|---|
| 6.3 | CVSS 4.0 | Medium | — | — | disclosure@vulncheck.com via NIST National Vulnerability Database |
| 6.3 | CVSS 4.0 | Medium | — | — | VulnCheck via CVE Program (cve.org) e CISA ADP |
| 4.8 | CVSS 3.1 | Medium | 2.2 | 2.5 | disclosure@vulncheck.com via NIST National Vulnerability Database |
| 4.8 | CVSS 3.1 | Medium | — | — | GitHub, Inc. via GitHub Security Advisories |
| 4.8 | CVSS 3.1 | Medium | — | — | VulnCheck via CVE Program (cve.org) e CISA ADP |
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:L/I:L/A:N (score shown at the top: disclosure@vulncheck.com) Weakness MITRE
Weakness class according to MITRE’s CWE catalog.
Detection
Read off the CVSS vector and the weakness class. Starting points, not tested rules.
- Prioritize edge telemetry for the network-reachable component (proxy, WAF, IDS/IPS).
Timeline
What happened to this CVE, newest first.
- 2026
- SSVC assessment updated
- Affected products added · sgl-project/sglang
- Record published
- CVSS changed · 4.8
- Source status changed · → Deferred
- Record created in this database
- CVE published
Source activity
Repeated readings are counted rather than listed.
- 2× References added · Oct 16, 2026 – Oct 16, 2026
- 1× Official description updated · Oct 16, 2026 – Oct 16, 2026
References
7 on the record
- github.com/sgl-project/sglang/pull/44653
- www.vulncheck.com/advisories/sglang-through-0.5.21-kv-cache-poisoning-via-mm-hashes-in-
- nvd.nist.gov/vuln/detail/CVE-2026-111392
- github.com/advisories/GHSA-fgff-m398-2gh5
- github.com/sgl-project/sglang/issues/44641
- github.com/sgl-project/sglang/blob/v0.5.21/python/sglang/srt/managers/tokenizer_
- github.com/sgl-project/sglang/blob/v0.5.21/python/sglang/srt/managers/io_struct.
Elsewhere on this site
Related CVEs (same products)
- High CVE-2026-111391 SGLang 0.5.10 through 0.5.22 Code Injection via Debug Dumper Filter 12h
- Medium CVE-2026-110809 sgl-project SGLang Session Cleanup session_controller.py Session.create_req denial of service 2d
- Medium CVE-2026-110807 sgl-project SGLang generate Endpoint io_struct.py denial of service 2d
- High CVE-2026-109228 SGLang through 0.5.21 Resource Budget Bypass via GLM-V Video Parameters 4d
- High CVE-2026-109229 SGLang through 0.5.21 Decompression Bomb DoS via NanoNemotronVLImageProcessor 4d
- High CVE-2026-109227 SGLang through 0.5.21 Memory Exhaustion DoS via Qwen-VL Video Preprocessing 4d
Not in any source we poll
Listed rather than left blank: an empty field and an unmeasured one look identical on screen, and only one is a reason to look elsewhere.
- No FIRST EPSS score.
- No fixed version, patch or mitigation documented.
- No confirmed IOCs, IP addresses, domains, file hashes, or malware artifacts supplied.
- No organization-specific asset inventory, compensating-control status, or patch deployment evidence supplied.
- No exploit packet captures, log samples, or incident case IDs supplied.
Answered from this record
Text assembled only from the fields above — no additional claims.
What should defenders know first?
CVE-2026-111392 is "SGLang through 0.5.21 KV Cache Poisoning via mm_hashes in /generate", rated Medium. It affects SGLang (Sgl-project); versions are in the scope table. It is not in CISA KEV, which does not rule out exploitation. CVSS 4.8 describes technical severity. First verify whether exposed or business-critical assets run the affected versions, apply vendor fixes or mitigations, and preserve logs for review.
Comments
Moderated technical discussion. Read the community guidelines.
No comments yet.
Sign in or create an account to comment and suggest corrections.